Jobs monitor

Is your feature request related to a problem? Please describe.

In my LocalAI use case, users send tasks through the Nextcloud Assistant and the OpenAI and LocalAI Integration.

The tasks:

• can be of different type: transcription, summaries, chatting .. and can require long time to complete.
• can come concurrently and from multiple instances so that I'd like to know how and when these tasks are executing, the resources they are using.

Currently I don't have a way to know the tasks currently running in LocalAI.

Describe the solution you'd like

It would be great to have a way to monitor the jobs that are sent to LocalAI. The way I think it would useful, is to have a Linux top-like command in which I can see information about running jobs:

• job id
• job status: executing, queued ...
• start time
• command
• backend
• model
• client (IP and / or user agent)

Other interesting features that could be considered:

• To have a way to kill a job.
• Be able to see past jobs with their start and end times.
• To have this monitoring panel in the Web UI

Describe alternatives you've considered
I had a look at the endpoint /backends/jobs but I get a "Resource not found" error (also reported here).

Additional context
I'm not sure if what I'm asking is already possible with current API, but as mentioned just above my attempts were unsuccessful.

Start by investigating the /backends/jobs endpoint and the existing LocalAI API behavior, including why it returns "Resource not found." Define how running and past jobs should expose job IDs, status, timing, command, backend, model, client, and optional cancellation before implementing monitoring or a Web UI.